The present invention relates to anti-myostatin antibodies and methods of using the same. The present invention also relates to polypeptides containing variant Fc regions and methods of using the same. Myostatin, also referred to as growth differentiation factor-8 (GDF8), is a secreted protein and is a member of the transforming growth factor-β (TGF-β) superfamily of proteins. Members of this superfamily possess growth-regulatory and morphogenetic properties (see, e.g., Kingsley et al., Genes Dev. 8(2):133-146 (1994), Hoodless et al., Curr. Top. Microbiol. Immunol. 228:235-272 (1998), and U.S. Pat. No. 5,827,733). Myostatin is expressed primarily in the developing and adult skeletal muscle and functions as a negative regulator of muscle growth. Systemic overexpression of myostatin in adult mice leads to muscle wasting (see, e.g., Zimmers et al., Science 296(5572):1486-1488 (2002)) while, conversely, a myostatin knockout mouse is characterized by hypertrophy and hyperplasia of the skeletal muscle resulting in two- to threefold greater muscle mass than their wild type littermates (see, e.g., McPherron et al., Nature 387(6628):83-90 (1997)).
Like other members of the TGF-β family, myostatin is synthesized as a large precursor protein containing an N-terminal propeptide domain, and a C-terminal domain considered as the active molecule (see, e.g., McPherron and Lee, Proc. Natl. Acad. Sci. USA 94(23):12457-12461 (1997): WO 1994/021681). Two molecules of myostatin precursor are covalently linked via a single disulfide bond present in the C-terminal growth factor domain. Active mature myostatin (disulfide-bonded homodimer consisting of the C-terminal growth factor domain) is liberated from myostatin precursor through multiple steps of proteolytic processing. In the first step of the myostatin activation pathway, a peptide bond between the N-terminal propeptide domain and the C-terminal growth factor domain, Arg266-Asp267, is cleaved by a furin-type proprotein convertase in both chains of the homodimeric precursor. But the resulting three peptides (two propeptides and one mature myostatin (i.e., a disulfide-bonded homodimer consisting of the growth factor domains)) remain associated, forming a noncovalent inactive complex that is referred to as “latent myostatin.” Mature myostatin can then be liberated from latent myostatin through degradation of the propeptide. Members of the bone morphogenetic protein 1 (BMP1) family of metalloproteinases cleave a single peptide bond within the propeptide, Arg98-Asp99, with concomitant release of mature, active myostatin, a homodimer (see, e.g., Szlama et al., FEBS J 280(16):3822-3839 (2013)). Moreover, the latent myostatin can be activated in vitro by dissociating the complex with either acid or heat treatment as well (see, e.g., Lee, PloS One 3(2):e1628 (2008)).
Myostatin exerts its effects through a transmembrane serine/threonine kinase heterotetramer receptor family, activation of which enhances receptor transphosphorylation, leading to the stimulation of serine/threonine kinase activity. It has been shown that the myostatin pathway involves an active myostatin dimer binding to the activin receptor type IIB (ActRIIB) with high affinity, which then recruits and activates the transphosphorylation of the low affinity receptor, the activin-like kinase 4 (ALK4) or activin-like kinase 5 (ALK5). It has also been shown that the proteins Smad 2 and Smad 3 are subsequently activated and form complexes with Smad 4, which are then translocated to the nucleus for the activation of target gene transcription. It has been demonstrated that ActRIIB is able to mediate the influence of myostatin in vivo, as expression of a dominant negative form of ActRIIB in mice mimics myostatin gene knockout (see, e.g., Lee, Proc. Natl. Acad. Sci. USA 98(16):9306-9311 (2001)).
A number of disorders or conditions are associated with muscle wasting (i.e., loss of or functional impairment of muscle tissue), such as muscular dystrophy (MD; including Duchenne muscular dystrophy), amyotrophic lateral sclerosis (ALS), muscle atrophy, organ atrophy, frailty, congestive obstructive pulmonary disease (COPD), sarcopenia, and cachexia resulting from cancer or other disorders, as well as renal disease, cardiac failure or disease, and liver disease. Patients will benefit from an increase in muscle mass and/or muscle strength; however, there are presently limited treatments available for these disorders. Thus, due to its role as a negative regulator of skeletal muscle growth, myostatin becomes a desirable target for therapeutic or prophylactic intervention for such disorders or conditions, or for monitoring the progression of such disorders or conditions. In particular, agents that inhibit the activity of myostatin may be therapeutically beneficial.
Inhibition of myostatin expression leads to both muscle hypertrophy and hyperplasia (McPherron et al., Nature 387(6628):83-90 (1997)). Myostatin negatively regulates muscle regeneration after injury and lack of myostatin in myostatin null mice results in accelerated muscle regeneration (see, e.g., McCroskery et al., J Cell Sci 118(15):3531-3541 (2005)). Anti-myostatin (GDF8) antibodies described in, e.g., U.S. Pat. Nos. 6,096,506, 7,261,893, 7,320,789, 7,807,159, and 7,888,486, and WO 2005/094446, WO 2007/047112, and WO 2010/070094 have been shown to bind to myostatin and inhibit myostatin activity in vitro and in vivo, including myostatin activity associated with the negative regulation of skeletal muscle mass. Myostatin-neutralizing antibodies increase body weight, skeletal muscle mass, and muscle size and strength in the skeletal muscle of wild type mice (see, e.g., Whittemore et al., Biochem. Biophys. Res. Commun. 300(4):965-971 (2003)) and the mdx mice, a model for muscular dystrophy (see, e.g., Bogdanovich et al., Nature 420(6914):418-421 (2002); Wagner., Ann. Neurol. 52(6):832-836 (2002)). However, these prior art antibodies are all specific for mature myostatin but not for latent myostatin, and the strategies described for inhibiting myostatin activity have utilized antibodies that can bind to and neutralize mature myostatin.
Antibodies are drawing attention as pharmaceuticals since they are highly stable in blood and have few side effects (see, e.g., Reichert et al., Nat. Biotechnol. 23:1073-1078 (2005) and Pavlou et al., Eur. J. Pharm. Biopharm. 59:389-396 (2005)). Almost all therapeutic antibodies currently on the market are antibodies of the human IgG1 subclass. One of the known functions of IgG class antibodies is antibody-dependent cell-mediated cytotoxicity (hereinafter denoted as ADCC activity) (see, e.g., Clark et al., Chem. Immunol. 65:88-110 (1997)). For an antibody to exhibit ADCC activity, the antibody Fc region must bind to an Fcγ receptor (hereinafter denoted as FcγR) which is an antibody-binding receptor present on the surface of effector cells such as killer cells, natural killer cells, and activated macrophages.
In humans, the FcγRIa (CD64A), FcγRIIa (CD32A), FcγRIIb (CD32B), FcγRIIIa (CD16A), and FcγRIIIb (CD16B) isoforms have been reported as the FcγR protein family, and the respective allotypes have also been reported (see, e.g., Jefferis et al., Immunol. Lett. 82:57-65 (2002)). FcγRIa, FcγRIIa, and FcγRIIIa are called activating FcγR since they have immunologically active functions, and FcγRIIb is called inhibitory FcγR since it has immunosuppressive functions (see, e.g., Smith et al., Nat. Rev. Immunol. 10:328-343 (2010)). In the binding between the Fc region and FcγR, several amino acid residues in the antibody hinge region and CH2 domain, and a sugar chain attached to Asn at position 297 (EU numbering) bound to the CH2 domain have been shown to be important (see, e.g., Radaev et al., J. Biol. Chem. 276:16478-16483 (2001), Greenwood et al., Eur. J. Immunol. 23:1098-1104 (1993), and Morgan et al., Immunology 86:319-324 (1995)). Various variants having FcγR-binding properties, mainly antibodies with mutations introduced into these sites, have been studied so far; and Fc region variants having higher binding activities towards activating FcγR have been obtained (see, e.g., WO 2000/042072, WO 2006/019447, WO 2004/099249, and WO 2004/029207).
When an activating FcγR is cross-linked with an immune complex, it phosphorylates immunoreceptor tyrosine-based activating motifs (ITAMs) contained in the intracellular domain or FcR common γ-chain (an interaction partner), activates a signal transducer SYK, and triggers an inflammatory immune response by initiating an activation signal cascade (see, e.g., Nimmerjahn et al., Nat. Rev. Immunol. 8:34-47 (2008)). FcγRIIb is the only FcγR expressed on B cells (see, e.g., Amigorena et al., Eur. J. Immunol. 19:1379-1385 (1989)). Interaction of the antibody Fc region with FcγRIIb has been reported to suppress the primary immune response of B cells (see, e.g., Sinclair, J. Exp. Med. 129:1183-1201 (1969)). Furthermore, it is reported that when FcγRIIb on B cells and B cell receptor (BCR) are cross-linked via an immune complex in blood, B cell activation and antibody production by B cells is suppressed (see, e.g., Heyman, Immunol. Lett. 88:157-161 (2003)). In this immunosuppressive signal transduction mediated by BCR and FcγRIIb, the immunoreceptor tyrosine-based inhibitory motif (ITIM) contained in the intracellular domain of FcγRIIb is necessary (see, e.g., Amigorena et al., Science 256:1808-1812 (1992) and Muta et al., Nature 368:70-73 (1994)). When ITIM is phosphorylated upon signaling, SH2-containing inositol polyphosphate 5-phosphatase (SHIP) is recruited, transduction of other activating FcγR signal cascades is inhibited, and inflammatory immune response is suppressed (see, e.g., Ravetch, Science 290:84-89 (2000)). Furthermore, aggregation of FcγRIIb alone has been reported to transiently suppress calcium influx due to BCR cross-linking and B cell proliferation in a BCR-independent manner without inducing apoptosis of IgM-producing B cells (see, e.g., Fournier et al., J. Immunol. 181:5350-5359 (2008)).
FcγRIIb is also expressed on dendritic cells, macrophages, activated neutrophils, mast cells, and basophils. FcγRIIb inhibits the functions of activating FcγR such as phagocytosis and release of inflammatory cytokines in these cells, and suppresses inflammatory immune responses (see, e.g., Smith et al., Nat. Rev. Immunol. 10:328-343 (2010)).
The importance of immunosuppressive functions of FcγRIIb has been elucidated so far through studies using FcγRIIb knockout mice. There are reports that in FcγRIIb knockout mice, humoral immunity is not appropriately regulated (see, e.g., J. Immunol. 163:618-622 (1999)), sensitivity towards collagen-induced arthritis (CIA) is increased (see, e.g., Yuasa et al., J. Exp. Med. 189:187-194 (1999)), lupus-like symptoms are presented, and Goodpasture's syndrome-like symptoms are presented (see, e.g., Nakamura et al., J. Exp. Med. 191:899-906 (2000)).
Additionally, regulatory inadequacy of FcγRIIb has been reported to be related to human autoimmune diseases. For example, the relationship between genetic polymorphism in the transmembrane region and promoter region of FcγRIIb, and the frequency of development of systemic lupus erythematosus (SLE) (see, e.g., Blank, Hum. Genet. 117:220-227 (2005), Olferiev et al., J. Biol. Chem. 282:1738-1746 (2007), Chen et al., Arthritis Rheum. 54:3908-3917 (2006), Floto et al., Nat. Med. 11:1056-1058 (2005), and Li et al., J. Immunol. 176:5321-5328 (2006)), and decrease of FcγRIIb expression on the surface of B cells in SLE patients (see, e.g., Mackay et al., J. Exp. Med. 203:2157-2164 (2006) and Yang et al., J. Immunol. 178:3272-3280 (2007)) have been reported.
From mouse models and clinical findings as such, FcγRIIb is considered to play the role of controlling autoimmune diseases and inflammatory diseases through involvement in particular with B cells, and it is a promising target molecule for controlling autoimmune diseases and inflammatory diseases.
IgG1, mainly used as a commercially available therapeutic antibody, is known to bind not only to FcγRIIb, but also strongly to activating FcγR (see, e.g., Bruhns et al., Blood 113:3716-3725 (2009)). It may be possible to develop therapeutic antibodies having greater immunosuppressive properties compared with those of IgG1, by utilizing an Fc region with enhanced FcγRIIb binding, or improved FcγRIIb-binding selectivity compared with activating FcγR. For example, it has been suggested that the use of an antibody having a variable region that binds to BCR and an Fc with enhanced FcγRIIb binding may inhibit B cell activation (see, e.g., Chu et al., Mol. Immunol. 45:3926-3933 (2008)). It has been reported that crosslinking FcγRIIb on B cells and IgE bound to a B-cell receptor suppresses differentiation of B cells into plasma cells, which as a result causes suppression of IgE production; and in human PBMC-transplanted mice, human IgG and IgM concentrations are maintained whereas the human IgE concentration is decreased (see, e.g., Chu et al., J. Allergy Clin. Immunol. 129:1102-1115 (2012)). Besides IgE, it has been reported that when FcγRIIB and CD79b which is a constituent molecule of a B-cell receptor complex are cross-linked by an antibody, B cell proliferation is suppressed in vitro, and arthritis symptoms are alleviated in the collagen arthritis model (see, e.g., Veri et al., Arthritis Rheum. 62:1933-1943 (2010)).
Besides B cells, it has been reported that crosslinking of FcεRI and FcγRIIb on mast cells using molecules, in which the Fc portion of an IgG with enhanced FcγRIIb binding is fused to the Fc portion of IgE that binds to an IgE receptor FcεRI, causes phosphorylation of FcγRIIb, thereby suppressing FcεRI-dependent calcium influx. This suggests that inhibition of degranulation via FcγRIIb stimulation is possible by enhancing FcγRIIb binding (see, e.g., Cemerski et al., Immunol. Lett. 143:34-43 (2012)).
Accordingly, an antibody having an Fc with improved FcγRIIb-binding activity is suggested to be promising as a therapeutic agent for inflammatory diseases such as an autoimmune disease.
Furthermore, it has been reported that activation of macrophages and dendritic cells via Toll-like receptor 4 due to LPS stimulation is suppressed in the presence of an antibody-antigen immune complex, and this effect is also suggested to be actions of the immune complex via FcγRIIb (see, e.g., Wenink et al., J. Immunol. 183:4509-4520 (2009) and Zhang et al., J. Immunol. 182:554-562 (2009)). Therefore, use of antibodies with enhanced FcγRIIb binding is expected to enable enhancement of TLR-mediated activation signal-suppressing actions; thus such antibodies have been suggested as being promising as therapeutic agents for inflammatory diseases such as autoimmune diseases.
Additionally, mutants with enhanced FcγRIIb binding have been suggested to be promising therapeutic agents for cancer, as well as therapeutic agents for inflammatory diseases such as autoimmune diseases. So far, FcγRIIb has been found to play an important role in the agonistic activity of agonist antibodies against the anti-TNF receptor superfamily. Specifically, it has been suggested that interaction with FcγRIIb is required for the agonistic activity of antibodies against CD40, DR4, DR5, CD30, and CD137, which are included in the TNF receptor family (see, e.g., Ravetch, Science 333:1030-1034 (2011), Wilson et al., Cancer Cell 19:101-113 (2011), Kohrt et al., J. Clin. Invest. 122:1066-1075 (2012), Xu et al., J. Immunol. 171:562-568 (2003), Zhang et al., Blood 108:705-710 (2006), Chuntharapai et al., J. Immunol. 166:4891-4898 (2001) and Ravetch et al., Proc. Natl. Acad. Sci. USA 109:10966-10971 (2012)). Ravetch, Science 333:1030-1034 (2011) shows that the use of antibodies with enhanced FcγRIIb binding enhances the anti-tumor effect of anti-CD40 antibodies. Accordingly, antibodies with enhanced FcγRIIb are expected to have an effect of enhancing agonistic activity of agonist antibodies including antibodies against the anti-TNF receptor superfamily.
In addition, it has been shown that cell proliferation is suppressed when using an antibody that recognizes Kit, a type of receptor tyrosine kinase (RTK), to crosslink FcγRIIb and Kit on Kit-expressing cells. Similar effects have been reported even in cases where this Kit is constitutively activated and has mutations that cause oncogenesis (see, e.g., Cemerski et al., Immunol. Lett. 143:28-33 (2002)). Therefore, it is expected that use of antibodies with enhanced FcγRIIb binding may enhance inhibitory effects on cells expressing RTK having constitutively activated mutations.
Antibodies having an Fc with improved FcγRIIb-binding activity have been reported (see, e.g., Chu et al., Mol. Immunol. 45:3926-3933 (2008)). In this Literature, FcγRIIb-binding activity was improved by adding alterations such as S267E/L328F, G236D/S267E, and S239D/S267E to an antibody Fc region. Among them, the antibody introduced with the S267E/L328F mutation most strongly binds to FcγRIIb, and maintains the same level of binding to FcγRIa and FcγRIIa type H in which a residue at position 131 of FcγRIIa is His as that of a naturally-occurring IgG1. However, another report shows that this alteration enhances the binding to FcγRIIa type R in which a residue at position 131 of FcγRIIa is Arg several hundred times to the same level of FcγRIIb binding, which means the FcγRIIb-binding selectivity is not improved in comparison with type-R FcγRIIa (see, e.g., US Appl. Publ. No. US2009/0136485).
Only the effect of enhancing FcγRIIa binding and not the enhancement of FcγRIIb binding is considered to have influence on cells such as platelets which express FcγRIIa but do not express FcγRIIb (see, e.g., Smith et al., Nat. Rev. Immunol. 10:328-343 (2010)). For example, the group of patients who were administered bevacizumab, an antibody against VEGF, is known to have an increased risk for thromboembolism (see, e.g., Scappaticci et al., J. Natl. Cancer. Inst. 99:1232-1239 (2007)). Furthermore, thromboembolism has been observed in a similar manner in clinical development tests of antibodies against the CD40 ligand, and the clinical study was discontinued (see, e.g., Arthritis Rheum. 48:719-727 (2003)). In both cases of these antibodies, later studies using animal models and such have suggested that the administered antibodies aggregate platelets via FcγRIIa binding on the platelets, and form blood clots (see, e.g., Meyer et al., J. Thromb. Haemost. 7:171-181 (2008) and Robles-Carrillo et al., J. Immunol. 185:1577-1583 (2010)). In systemic lupus erythematosus which is an autoimmune disease, platelets are activated via an FcγRIIa-dependent mechanism, and platelet activation has been reported to correlate with the severity of symptoms (see, e.g., Duffau et al., Sci. Transl. Med. 2:47ra63 (2010)). Administering an antibody with enhanced FcγRIIa binding to such patients who already have a high risk for developing thromboembolism will increase the risk for developing thromboembolism, thus is extremely dangerous.
Furthermore, antibodies with enhanced FcγRIIa binding have been reported to enhance macrophage-mediated antibody dependent cellular phagocytosis (ADCP) (see, e.g., Richards et al, Mol. Cancer Ther. 7:2517-2527 (2008)). When antigens to be bound by the antibodies are phagocytized by macrophages, antibodies themselves are considered to be also phagocytized at the same time. When antibodies are administered as pharmaceuticals, it is supposed that peptide fragments derived from the administered antibodies are likely to be also presented as an antigen, thereby increasing the risk of production of antibodies against therapeutic antibodies (anti-therapeutic antibodies). More specifically, enhancing FcγRIIa binding will increase the risk of production of antibodies against the therapeutic antibodies, and this will remarkably decrease their value as pharmaceuticals. Furthermore, FcγRIIb on dendritic cells have been suggested to contribute to peripheral tolerance by inhibiting dendritic cell activation caused by immune complexes formed between antigens and antibodies, or by suppressing antigen presentation to T cells via activating Fcγ receptors (see, e.g., Desai et al., J. Immunol. 178:6217-6226 (2007)). Since FcγRIIa is also expressed on dendritic cells, when antibodies having an Fc with enhanced selective binding to FcγRIIb are used as pharmaceuticals, antigens are not readily presented by dendritic cells and such due to enhanced selective binding to FcγRIIb, and risk of anti-drug antibody production can be relatively decreased. Such antibodies may be useful in that regard as well.
More specifically, the value as pharmaceuticals will be considerably reduced when FcγRIIa binding is enhanced, which leads to increased risk of thrombus formation via platelet aggregation and increased risk of anti-therapeutic antibody production due to an increased immunogenicity.
From such a viewpoint, the aforementioned Fc variant with enhanced FcγRIIb binding shows significantly enhanced type-R FcγRIIa binding compared with that of a naturally-occurring IgG1. Therefore, its value as a pharmaceutical for patients carrying type-R FcγRIIa is considerably reduced. Types H and R of FcγRIIa are observed in Caucasians and African-Americans with approximately the same frequency (see, e.g., Salmon et al., J. Clin. Invest. 97:1348-1354 (1996) and Manger et al., Arthritis Rheum. 41:1181-1189 (1998)). Therefore, when this Fc variant was used for treatment of autoimmune diseases, the number of patients who can safely use it while enjoying its effects as a pharmaceutical will be limited.
Furthermore, in dendritic cells deficient in FcγRIIb or dendritic cells in which the interaction between FcγRIIb and the antibody Fc portion is inhibited by an anti-FcγRIIb antibody, dendritic cells have been reported to mature (see, e.g., Boruchov et al., J. Clin. Invest. 115:2914-2923 (2005) and Dhodapkar et al., Proc. Natl. Acad. Sci. USA 102:2910-2915 (2005)). This report suggests that FcγRIIb is actively suppressing maturation of dendritic cells in a steady state where inflammation and such are not taking place and activation does not take place. FcγRIIa is expressed on the dendritic cell surface in addition to FcγRIIb; therefore, even if binding to inhibitory FcγRIIb is enhanced and if binding to activating FcγR such as FcγRIIa is also enhanced, maturation of dendritic cells may be promoted as a result. More specifically, improving not only the FcγRIIb-binding activity but also the ratio of FcγRIIb-binding activity relative to FcγRIIa-binding activity is considered to be important in providing antibodies with an immunosuppressive action.
Therefore, when considering the generation of a pharmaceutical that utilizes the FcγRIIb binding-mediated immunosuppressive action, there is a need for an Fc variant that not only has enhanced FcγRIIb-binding activity, but also has binding to both FcγRIIa types H and R allotypes, which is maintained at a similar level or is weakened to a lower level than that of a naturally-occurring IgG1.
Meanwhile, cases where amino acid alterations were introduced into the Fc region to increase the FcγRIIb-binding selectivity have been reported so far (see, e.g., Armour et al., Mol. Immunol. 40:585-593 (2003)). However, all variants said to have improved FcγRIIb selectivity as reported in this literature showed decreased FcγRIIb binding compared with that of a naturally-occurring IgG1. Therefore, it is considered to be difficult for these variants to actually induce an FcγRIIb-mediated immunosuppressive reaction more strongly than IgG1.
Furthermore, since FcγRIIb plays an important role in the agonist antibodies mentioned above, enhancing their binding activity is expected to enhance the agonistic activity. However, when FcγRIIa binding is similarly enhanced, unintended activities such as ADCC activity and ADCP activity will be exhibited, and this may cause side effects. Also from such viewpoint, it is preferable to be able to selectively enhance FcγRIIb-binding activity.
From these results, in producing therapeutic antibodies to be used for treating autoimmune diseases and cancer utilizing FcγRIIb, it is important that compared with those of a naturally-occurring IgG, the activities of binding to both FcγRIIa allotypes are maintained or decreased, and FcγRIIb binding is enhanced. However, FcγRIIb shares 93% sequence identity in the extracellular region with that of FcγRIIa which is one of the activating FcγRs, and they are very similar structurally. There are allotypes of FcγRIIa, H type and R type, in which the amino acid at position 131 is His (type H) or Arg (type R), and yet each of them reacts differently with the antibodies (see, e.g., Warmerdam et al., J. Exp. Med. 172:19-25 (1990)). Therefore, the difficult problem may be producing an Fc region variant with enhanced selective FcγRIIb binding as compared to each allotype of FcγRIIa, which involves distinguishing highly homologous sequences between FcγRIIa and FcγRIIb. In spite of those difficulties, several Fc region variants have been identified so far, which has selective binding activity to FcγRIIb as compared to FcγRIIa, by conducting comprehensive amino acid modification analysis in the Fc region (see, e.g., 2012/115241, WO 2013/047752, WO 2013/125667, WO 2014/030728 and WO 2014/163101).
There has been a report on an Fc region variant with binding selectivity for FcγRIIb in relation to human FcγR so far, whereas there has been no report on an Fc region variant with binding selectivity for FcγRIIb in relation to monkey FcγR. Owing to the absence of such an Fc variant, the effects of the Fc variant selectively binding to FcγRIIb have not been thoroughly tested yet in monkey.
Apart from the above, it is reported that by modifying the charge of amino acid residues which may be exposed on the surface of an antibody so as to increase or decrease the isoelectric point (pI) of the antibody, it is possible to regulate the half-life of the antibody in blood (see, e.g., WO 2007/114319 and WO 2009/041643). They show that it is possible to prolong the plasma half-life of an antibody by reducing the antibody's pI and vice versa.
Further, it is reported that incorporation of an antigen into cells can be promoted by modifying the charge of specified amino acid residues particularly in its CH3 domain to increase the antibody's pI (see, e.g., WO 2014/145159). Also, it has been reported that modifying the charge of amino acid residues in the constant region (mainly CH1 domain) of an antibody to reduce pI can prolong the half-life of the antibody in plasma (see, e.g., WO 2012/016227).